Review of Sasaki et al., 2021 GMDD

The authors present an offline simulation of dissolved and particulate Pa and Th, and sedimentary Pa/Th. The effect of bottom scavenging and scavenging efficiency depending on particle concentration are discussed through sensitivity experiments. The description of the modeling results are in great detail and the authors state that the model reproduces the observations reasonably well. However, there are still many places of model-data discrepancy. I understand that it is impossible to perfectly reproduce the observation, but the mismatch should at least be discussed. In addition, this study lacks novelty. It seems that it is confirming previous other modeling results and provides limited new results for improving our understanding of marine Pa and Th cycle. Also, a critical comparison of this modeling results of sedimentary Pa/Th with other modeling studies is missing.

Detailed comments are listed below:

In the introduction, the authors reviewed previous modeling studies. From the methods part, I see this study is offline modeling, which is a big difference from other modeling studies and should be emphasized.

Line 100, why not using observations for calcite and opal? Why calcite and opal export productions are calculated using ratio and POC observation?

Line 154, sensitivity experiments in this study is using off-line model, it is not appropriate to call them “OGCM experiments”, also later in the text.

What is the thickness of the nepheloid layer in the model?

Rempfer et al.,(2017) includes the bottom scavenging in their model, how this bottom scavenging implementation in this study differ from Rempfer et al., (2017)?

Line 170, C_total and C_ref, are they global average or values in each grid? Is C_total on each grid and C_ref a global mean?

Figure 3, although with bottom scavenging, the Pa_d concentration decreases below ~3,000m, from the vertical profile, the observation shows similar values below 3,000m with no decreasing trend with depth, but the model shows a clear decreasing trend (Figure 3d). How to explain this model-data discrepancy? It seems that the bottom scavenging is too strong.

Line 221, the authors pointed out the lower than observation Th_d, what could potentially cause this mismatch? Can it be improved in this model?

Line 250-251 “Similar features are also found for dissolved 230Th”. For Pa_d, the maximum around 3,000 is more or less reproduced. But for Th, the maximum in the model is much deeper than the observation. Near 30W, the simulated mid-depth Th_d is much lower than the observation. Also in Figure S6.

Line 251-252, the authors state that the particulate Pa and Th are well reproduced. It is not obvious in Figure S5c and d due to the colormap scale. In Figure S5c, observations are all in dark blue (hard to tell the value from the color bar), but simulation has some green-yellow values. What processes in the model cause these high values in Pa_P? Also in Figure S5, Pa_P/Th_P should also be compared with observation.

Line 305, results about 1D_EXP. Sedimentary Pa/Th in the Atlantic is influenced by AMOC and particle flux effect in Siddall et al., (2005), probably AMOC is the first order factor (Gu & Liu, 2017). In this study, transport includes both of these effect. More experiments can be carried out to separate the effect of ocean currents and the diffusion caused by particle effects. In this off-line model, it is computationally achievable.

How is simulated sedimentary Pa/Th in CTRL and 3D_EXP compare with observation quantitatively? In Rempfer et al., (2017), the bottom scavenging is suggested to not affect Pa_P/Th_P to a small extent and also not affect the relationship between Pa_P/Th_P and AMOC. Does this study support the results in Rempfer et al., (2017)?

Also, how is the sedimentary Pa/Th in this study compare quantitatively with other previous models? This is important because sedimentary Pa/Th is an important paleo proxy.

Line 334-336 is confusing. Please rephrase to make it more clear.

Line 339-340, what is the “ocean transport” here mean? Southward transport by the lower limb of Atlantic Meridional Overturning Circulation or diffusion?

Line 340-342,”At the same time…; as a result… ” I cannot follow the logic here.

Line 379: Residence time is calculated in CTRL_EXP and Siddall_EXP, with bottom scavenging the residence time is significantly decreased. However, the residence time in CTRL_EXP is similar to the residence time in (Gu & Liu, 2017) which does not include bottom scavenging. Does this mean the correct residence time does not necessarily need bottom scavenging?

Line 426 “Part of the error comes from the oceanic flow fields simulated in the ocean model”. How is the oceanic flow simulated in this model? It can be verified against products such as ECCO (Fukumori et al., 2018). Since Atlantic sedimentary Pa_P/Th_P is greatly influenced by AMOC, what does AMOC in this model look like? This information should be provided in the manuscript.

Line 433-434, why not in this study? Results presented in this study are mostly confirmation of previous studies. With the efficiency of this off-line model, more things can be done for example test this particle fields effect on Pa and Th.

The summary is too verbose. Line 435-480 is repeating things in the previous section.

Line 486-489 simulated sedimentary Pa/Th under glacial times are also discussed in a 2-D model (Lippold et al., 2012) and recently in a 3-D model (Gu et al., 2020).

Quantitative model data agreement (Pa_D, Th_D, Pa_P, TH_P, and sedimentary Pa/Th) and also residence time can be summarized in a table for different experiments. In that way, the performance of different experiments can be clearly seen.

Minor issues:

Colorbars only have the highest and lowest values, it is not easy to tell the values in the middle.

Line 38, particulate organic carbon.

Line 47, Gu & Liu, (2017) also show AMOC change on sedimentary Pa/Th. Also, in Gu & Liu, (2017), the particle change due to freshwater and its impact on sedimentary Pa/Th is examined and should be cited in line 432.

Figure 6e, a lot of observations overlapped. It is hard to tell the color.

Figure 4 can be plotted in one figure as Figure 4 in Rempfer et al., (2017). In this way, the relative difference between different experiments is clearly shown.

References:

Fukumori, I., Heimbach, P., Ponte, R. M., & Wunsch, C. (2018). A dynamically consistent, multivariable ocean climatology. Bulletin of the American Meteorological Society, 99(10), 2107–2127. https://doi.org/10.1175/BAMS-D-17-0213.1

Gu, S., & Liu, Z. (2017). 231Pa and 230Th in the ocean model of the Community Earth System Model (CESM1.3 ). Geoscientific Model Development, 10, 4723–4742. https://doi.org/10.5194/gmd-10-4723-2017

Gu, S., Liu, Z., Oppo, D. W., Lynch-Stieglitz, J., Jahn, A., Zhang, J., & Wu, L. (2020). Assessing the potential capability of reconstructing glacial Atlantic water masses and AMOC using multiple proxies in CESM. Earth and Planetary Science Letters, 541, 116294. https://doi.org/10.1016/j.epsl.2020.116294

Lippold, J., Luo, Y., Francois, R., Allen, S. E., Gherardi, J., Pichat, S., et al. (2012). Strength and geometry of the glacial Atlantic Meridional Overturning Circulation. Nature Geoscience, 5(11), 813–816. https://doi.org/10.1038/ngeo1608

Rempfer, J., Stocker, T. F., Joos, F., Lippold, J., & Jaccard, S. L. (2017). New insights into cycling of 231 Pa and 230 Th in the Atlantic Ocean. Earth and Planetary Science Letters, 468, 27–37. https://doi.org/10.1016/j.epsl.2017.03.027

Siddall, M., Henderson, G. M., Edwards, N. R., Frank, M., Müller, S. a., Stocker, T. F., & Joos, F. (2005). 231Pa/230Th fractionation by ocean transport, biogenic particle flux and particle type. Earth and Planetary Science Letters, 237(1–2), 135–155. https://doi.org/10.1016/j.epsl.2005.05.031

This study explores the processes that control the distribution of 231Pa and 230Th in the oceans and underlying sediments using COCO V4.0, an Ocean General Circulation Model (OGCM), from Hasumi 2006.

They implemented 231Pa and 230Th in the model using offline tracer simulations based on physical fields from COCO. They implemented bottom scavenging as well as a “dependence of scavenging efficiency on particle concentration” in the model.

General comments

- The most puzzling aspect of this manuscript is the lack of use of recent modeling results and the almost total lack of comparison with these model simulations (e.g. Missiaen et al., 2020a and 2020b; van Hulten et al., 2018; Rempfer et al., 2017; Lippold et al., 2012; Luo et al., 2010; Dutay et al., 2009; Roy-Barman, 2009). It is all the more surprising that most of these papers are cited by the authors although mostly as examples of recent publications instead of being analyzed in depth and compared to the COCO model outputs. A more thorough assessment of these new model simulations and how / why they agree / differ from the simulations presented in the manuscript must be done before publication.

- Similarly, the choice of comparing the COCO model outputs with those of one of the earliest models used for 231Pa and 230Th, namely the model from the Siddall et al., 2005 study, is very disappointing as it misses out all the improvements made by the newer modeling studies and most of the conclusions drawn from these more recent simulations, most of which representing a significant improvement from the Siddall et al. (2005) model. The authors need to carefully and thoroughly justify their choice. Nevertheless, an in-depth discussion to compare their model outputs and conclusions with that of the more recent modeling studies is needed and should not be limited, as it is the case in the present manuscript, to a comparison with the Siddall at al. (2005) simulation.

- In the same vein, there is a great lack of recent literature analysis on 231Pa and 230Th, e.g. the recent review by Costa et al. (2020) or the recent findings of Missiaen et al. (2018) on the effect of the detrital (238U/232Th) activity ratio on the calculation of 231Pa and 230Thxs are neither discussed or cited. A lot of the effects that the authors are discussing in their manuscript is actually discussed in details for 230Th in the review paper by Costa et al. (2020).

- The literature used to discuss the effect of particles type and distribution is neither the first/pioneering papers on the topics nor the latest. The authors should read the review by Costa et al. (2020) and look at the modeling results of Missiaen et al. (2020b) and references therein. These results should be both mentioned in the state-of-the-art section of the Introduction and later discussed.

- Similarly, the older literature is fundamentally overlooked. The term “boundary scavenging” has been defined and used by Anderson et al. (1983b). Part of what the authors seem to define as a discovery on the effect of particle concentration on scavenging is actually perfectly defined and modeled by Anderson and co-authors is this paper and subsequent papers. This leads to a conceptual problem L356-369 (see also comment on L194 below).

- Several sentences or model presentation are very vague, e.g. in equation 4a, there is a term “Transport” (L116-120) defined as representing transport by advection, diffusion and convection. These are 3 very distinct physical processes in their formulation, why is the term “Transport” not explicitly given? What does the term “convection” represent in the oceans. There is no bottom heating so I have great troubles understanding what the authors mean here.

- I am very puzzled by the use of equation (10) (L169) for both 230Th and 231Pa. The partition coefficient cannot be the same for both radionuclides as they as have different behaviors. The value of the exponent used here (-0.42) has been given by Henderson et al. (1999) for 230Th and is indeed not valid for 231Pa. I do not see what can be achieved by using the same reference partition coefficient for both isotopes.

- L90: there is one class of settling velocity in the model presented here. There are two classes in van Hulten et al. (2018). Since the authors discuss the effect of the concentration of particles on scavenging, they should discuss the effect of having one vs. more classes of settling speed on their conclusions

- L194: The authors say they included bottom scavenging in benthic nepheloid layers. This is a very important aspect of the model. However, how this is done is not explained. More explanations of this very important aspect are necessary, especially considering the objective of the journal.

- Amongst the conclusions, some are included in the equation. The fact that 231Pa is more affected by advection is 1) the basis for using Pa/Th as a proxy for ocean circulation and has already been verified by several models, and 2) is somehow imbedded in the equations of scavenging.

- English should be proofread. The meaning of several sentences remains very ambiguous or unclear.

To conclude on these general comments: the model and its interpretations seems detached from what is already known on Pa/Th both in the water column and the sediment from both modeling and data studies. This manuscript shows a lack of thorough reading (state-of-the-art) of the most recent (last 10 years) literature on the subject and lacks discussion of these recent findings / conclusions. The choice of using one of the oldest model to compare these new simulation results is very odd and thus lacks a great part of the novelty added by more recent studies. There are also several conceptual problems that need to be adressed.

Specific comments

- L24-25: if one wants to cover the all date range, there are more recent papers than Bohm et al. (2015), e.g. Sufke et al., 2020 or Waelbroeck et al. 2018

- L30: there is also Henderson and Anderson 2003 review that gives a large range of residence times (see also Costa et al., 2020 for 230Th)

- L36: for the LGM/Holocene comparison, there are more appropriate references, such as Lippold et al., 2014 which is a modeling and compilation of Atlantic data for the LGM vs. Holocene.

- L44 and after: several references missing or not cited appropriately. Many of the references cited cover several aspects of the Pa/Th modeling rather than only a specific aspect as the citation format made by the authors suggests.

- L64: GEOTRACE database: cite

- L76: 43 vertical layers: are these of uniform or different heights. Be more precise.

- L81: how do you assess that you reached a steady state? explain

- L81: Explain why you choose 100 years average rather than another number

- L171: “reference concentration”. It is very unclear to me, based on the information given here what is the reference concentration. More details should be given.

References:

Anderson R.F., Bacon M.P., and Brewer P.G. (1983b) Removal of ²³⁰Th and ²³¹Pa from the ocean margins. Earth Planet. Sci. Lett. 66, 73-90

Costa K.M., Hayes C.T., Anderson R.F., Pavia F.J., Bausch A., Deng F., Dutay J.-C., Geibert W., Heinze C., Henderson G., Hillaire-Marcel C., Hoffmann S., Jaccard S.L., Jacobel A.W., Kienast S.S., Kipp L., Lerner P., Lippold J., Lund D., Marcantonio F., McGee D., McManus J.F., Mekik F., Middleton J.L., Missiaen L., Not C., Pichat S., Robinson L.F., Rowland G.H., Roy-Barman M., Tagliabue A., Torfstein A., Winckler G., and Zhou Y. (2020) ²³⁰Th Normalization: New Insights on an Essential Tool for Quantifying Sedimentary Fluxes in the Modern and Quaternary Ocean. Paleoceanogr. Paleoclim. 35, e2019PA003820, 10.1029/2019pa003820.

Lippold J., Luo Y., Francois R., Allen S.E., Gherardi J., Pichat S., Hickey B., and Schulz H. (2012) Strength and geometry of the glacial Atlantic Meridional Overturning Circulation. Nature Geosci 5, 813-816, http://www.nature.com/ngeo/journal/v5/n11/abs/ngeo1608.html#supplementary-information

Missiaen L., Pichat S., Waelbroeck C., Douville E., Bordier L., Dapoigny A., Thil F., Foliot L., and Wacker L. (2018) Downcore Variations of Sedimentary Detrital (238U/232Th) Ratio: Implications on the Use of 230Thxs and 231Paxs to Reconstruct Sediment Flux and Ocean Circulation. Geochem. Geophys. Geosyst. 19, 2560-2573, https://doi.org/10.1029/2017GC007410.

Sufke F., Schulz H., Scheen J., Szidat S., Regelous M., Blaser P., Poppelmeier F., Goepfert T.J., Stocker T.F., and Lippold J. (2020) Inverse response of Pa-231/Th-230 to variations of the Atlantic meridional overturning circulation in the North Atlantic intermediate water. Geo-Mar. Lett. 40, 75-87, 10.1007/s00367-019-00634-7.

Waelbroeck C., Pichat S., Böhm E., Lougheed B.C., Faranda D., Vrac M., Missiaen L., Vazquez Riveiros N., Burckel P., Lippold J., Arz H.W., Dokken T., Thil F., and Dapoigny A. (2018) Relative timing of precipitation and ocean circulation changes in the western equatorial Atlantic over the last 45 kyr. Clim. Past 14, 1315-1330, 10.5194/cp-14-1315-2018.